Edge Computing Based Machine Learning Mobile Malware Detection

Huntsville (Ala.)

Because network infrastructure and bandwidth cannot keep up with the increasing demands, edge computing has sprung up to reorganize the cloud computing paradigm. As an application of edge computing, mobile malware detection requires significant resources that are unavailable to individual devices. The immediacy of malware detection is highly pertinent, as malicious software may already be in operation. In addition, the dispersion of mobile devices and massive simultaneous traffic make the offloading this work to remote data centers infeasible. In our study, we implement a typical edge node to perform Android malware detection. In our study, we collect an exceptional amount of mobile malware and benign samples, and demonstrate the ability of edge computing nodes to provide typical security functions for edge devices. We also show the ability to scale the operation for increasing data volumes.